What is web analytics?
Web analytics answers the questions every site owner eventually asks: how many people came, where from, what they looked at, what they did and where they left. A small script on each page, or the logs on the server, records visits and actions, and a tool turns them into reports: traffic by source, popular pages, devices, locations, conversions and the paths between them. Google Analytics is the most widely used; Adobe Analytics serves large enterprises; privacy-focused tools such as Plausible, Fathom and Matomo collect less and keep it simpler.
Modern analytics is built on events rather than page views. Every meaningful action, a video play, a scroll to the end, a form submission, a download, a purchase, is tracked as an event with details attached, and the most important ones are marked as conversions. Google made that model the default when Google Analytics 4 replaced its predecessor in 2023, which made measurement more flexible and more dependent on planning what to track.
Why it matters
Why web analytics matters
Without it, every decision about a website is an opinion. Analytics shows which campaigns bring visitors who act, which pages lose people, which content earns its place and which device is quietly underperforming. It turns a redesign from a matter of taste into a response to evidence, and it lets a brand show what its marketing returned. It also has limits that matter: consent rules, ad blockers and browser privacy features mean a share of visits is never recorded, so the numbers are a strong guide rather than a complete count.
In practice
How web analytics works in practice
- 01
Good analytics is planned before it is installed. We start with the business questions, what counts as success for this site, then write a measurement plan: the conversions, the events that lead to them, the details each needs and the naming conventions that keep reports readable. Tracking is installed, usually through a tag manager, consent is handled properly, and every event is tested before launch so the first month of data can be trusted.
- 02
Then the data is read regularly, not just collected. A short dashboard covers the few numbers that matter and is reviewed on a set rhythm, with deeper analysis when something moves. Analytics is paired with search data from Search Console and with qualitative tools such as heatmaps and session recordings, because numbers say what happened and rarely why.
Watch out
What people get wrong about web analytics
Collecting everything and analyzing nothing, so the account fills with events nobody looks at. Second, worshipping vanity metrics: page views and follower counts that rise without any change in sales, leads or sign-ups. Third, trusting numbers that were never tested. Duplicate tags, broken consent settings and unfiltered internal traffic quietly corrupt the data, and decisions made on bad data are worse than guesses because they feel certain. And reading a single week in isolation, when seasonality, a holiday or one newsletter can move any number on its own.
